Box Filter

Overview

Box Filter is a low-pass filter that smooths the image by making each output pixel the average of the surrounding ones, removing details, noise and and edges from images.

Input	Window size	Output
	5x5

Implementation

The Box filter is implemented as a convolution operation on the input image using the following kernel:

\[ box_{m,n} = \frac{1}{mn} \begin{bmatrix} 1 & 1 & \dots & 1 \\ 1 & 1 & \dots & 1 \\ \vdots & \vdots & \ddots & \vdots \\ 1 & 1 & \dots & 1 \end{bmatrix}_{m \times n} \]

Usage

Language: C/C++ Python

1. Import VPI module

   import vpi

2. Use the CUDA backend to filter input image with a 5x5 box kernel, using ZERO boundary condition. Input and output are VPI images.

   with vpi.Backend.CUDA:
       output = input.box_filter(5, border=vpi.Border.ZERO)

1. Initialization phase
   1. Include the header that defines the box filter function.

      #include <vpi/algo/BoxFilter.h>

   2. Define the input image object.

          VPIImage input = /*...*/;

   3. Create the output image. It gets its dimensions and format from the input image.

          int32_t w, h;
          vpiImageGetSize(input, &w, &h);
       
          VPIImageFormat type;
          vpiImageGetFormat(input, &type);
       
          VPIImage output;
          vpiImageCreate(w, h, type, 0, &output);

   4. Create the stream where the algorithm will be submitted for execution.

          VPIStream stream;
          vpiStreamCreate(0, &stream);

2. Processing phase
   1. Submit the algorithm to the stream along with other parameters. In this case it's a 5x5 box filter with clamp border extension. It'll be executed by the CUDA backend.

          vpiSubmitBoxFilter(stream, VPI_BACKEND_CUDA, input, output, 5, 5, VPI_BORDER_ZERO);

   2. Optionally, wait until the processing is done.

          vpiStreamSync(stream);

3. Cleanup phase
   1. Free resources held by the stream and the input and output images.

          vpiStreamDestroy(stream);
          vpiImageDestroy(input);
          vpiImageDestroy(output);

For more information, see Box Filter in the "API Reference" section of VPI - Vision Programming Interface.

Limitations and Constraints

Constraints for specific backends supersede the ones specified for all backends.

All Backends

• Kernel size must be at least 1x1 and at most 11x11.
• Accepted border extensions:
  • VPI_BORDER_CLAMP
  • VPI_BORDER_ZERO
• Input and output images must have same image type and size.
• Accepted image formats:
  • VPI_IMAGE_FORMAT_U8
  • VPI_IMAGE_FORMAT_S8
  • VPI_IMAGE_FORMAT_U16
  • VPI_IMAGE_FORMAT_S16

PVA

• Only available on Jetson Xavier devices.
• Minimum kernel size is 2x2.
• Accepted border extensions:
  • VPI_BORDER_ZERO
  • VPI_BORDER_CLAMP only if kernel size is 3x3.

VIC

• Not implemented.

Performance

For information on how to use the performance table below, see Algorithm Performance Tables.
Before comparing measurements, consult Comparing Algorithm Elapsed Times.
For further information on how performance was benchmarked, see Performance Benchmark.

clear filters Device: Jetson AGX Xavier Jetson TX2 Jetson Nano  -  Streams: 1 4 8